I'm trying out CentOS 7 using a HP AMD laptop http://h10025.www1.hp.com/ewfrf/wc/document?cc=uk&lc=en&docname=c038...
The problem is the laptop doesn't suspend properly in runlevel 5 / graphical.target and following guides, I found that it suspends (power LED blinking) but does not resume in runlevel 3 / multi-user.target. When it locks up, the machine is unresponsive to ping/ssh so it's not just a blank screen.
Unfortunately, the only applicable guide I found on troubleshooting this is for Ubuntu, which requires the kernel to support pm_trace which isn't found in /sys/power.
Since there is a big red warning on centos.org about building/compiling my own kernel to add functions, I'm wondering if there are any alternative method to troubleshoot the issue on CentOS.
I suspect it may be related to the Radeon GPU as I notice that every now and then after fresh boot, the GUI desktop will freeze for a while and dmesg will contain errors: [drm:cik_ib_test] *ERROR* radeon: fence wait failed (-35) [drm:radeon_ib_ring_tests] *ERROR* radeon: failed testing IB on ring 2 (-35)
As a side note, this laptop had no problems with suspend/resume when it was running Ubuntu 14.04
On 07/15/2014 11:16 AM, Emmanuel Noobadmin wrote:
I'm trying out CentOS 7 using a HP AMD laptop http://h10025.www1.hp.com/ewfrf/wc/document?cc=uk&lc=en&docname=c038...
The problem is the laptop doesn't suspend properly in runlevel 5 / graphical.target and following guides, I found that it suspends (power LED blinking) but does not resume in runlevel 3 / multi-user.target. When it locks up, the machine is unresponsive to ping/ssh so it's not just a blank screen.
Unfortunately, the only applicable guide I found on troubleshooting this is for Ubuntu, which requires the kernel to support pm_trace which isn't found in /sys/power.
Since there is a big red warning on centos.org about building/compiling my own kernel to add functions, I'm wondering if there are any alternative method to troubleshoot the issue on CentOS.
file an RFE for the centos-plus kernel, maybe we can get this in there ? Otherwise, nothing really stops you from building your own kernel for your own machine :)
On 7/16/14, Karanbir Singh mail-lists@karan.org wrote:
file an RFE for the centos-plus kernel, maybe we can get this in there ? Otherwise, nothing really stops you from building your own kernel for your own machine :)
Filed.
In the meantime, I'm going to try the suggestion of using Catalyst drivers first before venturing into kernel building since the laptop's graphics is behaving oddly apart from the resume problem.
In the meantime, I'm going to try the suggestion of using Catalyst drivers first before venturing into kernel building since the laptop's graphics is behaving oddly apart from the resume problem.
As an update, downloading the latest AMD fglrx 14.4rev2 drivers and installing it fixed the unable to resume problem to some extend. pm-hibernate will work properly but pm-suspend will wake up to a blank screen which apparently is a common enough problem from googling. Might be something to do with systemd based on this archlinux thread https://bbs.archlinux.org/viewtopic.php?id=178450
However, in my case, this appears to be circumvented in runlevel 5 (runlevel 3 gets stuck until I type blind or ssh in to reboot/hibernate the machine). In runlevel 5, the screen flashes on briefly, then go blank, after a while it comes back and abrt reports a crash with plymouthd. Seems like X will retry in this situation so it causes the display to come back.
If it is of any use, I could send the abrt report somewhere.
-
Emmanuel Noobadmin -
Karanbir Singh